Thirty states now have medical marijuana laws (MML), 9 have recreational marijuana laws (RML), and many more states are considering such laws. The health effects of cannabis laws are controversial; understanding them is a major public health and NIDA priority (PA-17-135). Thus far, only 3 studies of adults (2 of them ours; Hasin et al., 2017, Martins et al., 2016) used multi-level modeling to examine MML effects on cannabis outcomes with individual data. These studies suggested post-MML increases in cannabis use and Cannabis Use Disorder (CUD). However, they left many questions unanswered, including whether MML have stronger effects in those with key vulnerability factors (chronic pain, psychiatric disorders). In addition, soaring rates of opioid prescriptions and overdoses have led to calls for MML as part of the solution to the US opioid crisis, but most MML-opioid studies are ecological (a weak design to study individual behavior), and results from individual-level studies leave the evidence unclear. Ecological studies also suggest that through cannabis substitution, MML reduce medication prescriptions for common psychiatric disorders (e.g., PTSD, depression), but no individual-level studies of this have been conducted. Importantly, RML effects are almost entirely unknown, a major gap in knowledge. In Veterans Administration (VA) patients, CUD prevalence has doubled since 2002, and in those age ?35, is 2-6 times higher than in the general population. VA patients also have high rates of opioid prescriptions, overdoses, and of chronic pain and psychiatric disorders that may increase their vulnerability to adverse MML and RML effects. They thus are a large, vulnerable population in whom MML and RML effects are unknown. We will investigate MML and RML effects utilizing a major resource, the individual data from the VA Electronic Medical Record, available since 2000 from the ~5,000,000 patients served each year by the VA healthcare system. We will create yearly EMR datasets, and merge this with National Death Index data, Medicare data (for those age ?65) and state-year MML and RML variables that we will create. Using multi-level models and difference-in-difference tests, we will examine MML and RML effects on three main outcomes: cannabis (use, CUD), opioids (prescriptions, fatal and non-fatal overdoses, opioid use disorders), and psychotropic medication prescriptions (antidepressants, anxiolytics, sedatives/hypnotics). Importantly, we will determine if pain, psychiatric disorders or demographics (sex, age, race/ethnicity) modify MML/RML effects. We will also examine specific MML/RML provisions, time lags, and breakpoints in trends reflecting federal policy changes. Analyses will incorporate individual- and state-level confounders, e.g., state norms, economic factors. We will also explore alcohol and tobacco outcomes. The research team includes substance epidemiology/policy experts and VA addiction and internal medicine experts. Findings will be disseminated to clinicians and policy-makers. Determining MML/RML effects in VA patients will make a major contribution to the limited adult literature on MML/RML, contributing to knowledge that will inform policy and the care of individuals with vulnerability factors.